A Programming Model for Portable Fault Detection and Diagnosis

Abstract

Portable applications support the write once, deploy everywhere paradigm. This paradigm is particularly attractive in building applications, where current practice involves the manual deployment and configuration of such applications, requiring significant engineering effort and concomitant costs. This is a tedious and error-prone process which does not scale well. Notwithstanding recent advances in semantic data modelling that allow a unified representation of buildings, we still miss a paradigm for deploying portable building applications at scale. This paper introduces a portable programming model for such applications, which we examine in the context of Fault-Detection and Diagnosis (FDD). In particular, we look at the separation of the FDD logic and the configuration with specific data inputs. We architect a software system that enables their self-configuration and execution across various building configurations, expressed in terms of Brick metadata models. Our initial results from authoring and executing APAR (AHU Performance Assessment Rules) on multiple AHUs of two museums demonstrate the potential of our model to reduce repetitive tasks and deployment costs of FDD applications.